UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio and TV Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > Specific Vintage Equipment > Television Standards Converters, Modulators etc

Notices

Television Standards Converters, Modulators etc Standards converters, modulators anything else for providing signals to vintage televisions.

Closed Thread
 
Thread Tools
Old 9th Jul 2004, 3:04 pm   #1
tubesrule
Hexode
 
tubesrule's Avatar
 
Join Date: Jul 2004
Location: Michigan USA
Posts: 325
Default New Multi-Standards Converter / Convertor Design Discussion

Hello all,
I am posting here and elsewhere to determine the interest level in a Multi-Standards video converter I designed. I originally made these for myself and a few other collectors, but got some interest at the Early Television Convention this year, so I am trying to determine if it is feasible to have a few more built depending on if enough people are interested. I threw together a quick web page explaining the device and all the details. Please check it out at http://converter.home.comcast.net/ if you have any interest.

Thanks,
Darryl

(Modified to make the URL a clickable link by Paul S)

Last edited by Dave Moll; 4th Jun 2007 at 2:54 pm. Reason: update URL link
tubesrule is offline  
Old 9th Jul 2004, 8:41 pm   #2
ppppenguin
Retired Dormant Member
 
ppppenguin's Avatar
 
Join Date: Dec 2003
Location: North London, UK.
Posts: 6,168
Default Re: New Multi-Standards Converter

Fableglade: One small point to remember when running American sets. If they are 525 line sets the line scan frequency is very close to that used for 625 (15734Hz vs 15625Hz) so the line timebase will always lock to a 625 signal. There should be no difficulty getting the field timebase to lock to 50Hz instead of 60Hz. There are other minor differences between the 625/50 and 525/60 waveforms but they don't matter in this context. The only other snag I can think of It is possible that the higher frequency components in a 625 signal will cause buzz on sound with a USA TV.

All you have to worry about is modulating your video on to a suitable VHF carrier and doing sound at the correct spacing. Modern modulators as used in NTSC VCRs should be fine for this.

Darryl: That's an impressive little unit you have designed. I work professionally with Xilinx FPGAs etc and can't see how you can make and sell a box like that for £300, assuming that the price mentioned in one of the other messages is correct. That barely covers all the components and a PCB, let alone assembly and test.

ppppenguin is offline  
Old 9th Jul 2004, 9:27 pm   #3
tubesrule
Hexode
 
tubesrule's Avatar
 
Join Date: Jul 2004
Location: Michigan USA
Posts: 325
Default Re: New Multi-Standards Converter

Fableglade: That's the same reason I wanted to do this for so long, running 405 sets over here, and of course the mechanical sets. While you can usually modify the sets to the other scan rates, I like to keep them unmodified. It can however be a bit of an issue running the sets off the wrong main's frequency. Over here, running a 50Hz set at 60Hz isn't a problem for the power supply, but you can get hum bars in the video if it is not filtered well enough or there are stray magnetic fields near by. Over there you also need to make sure any of our transformer operated sets are rated for 50Hz main's or they may overheat. As ppppenguin pointed out, you can usually get the more modern sets to work on other standards with a little bit of work. I was actually running some 405 sets on 525 for a while, and while they wouldn't fill the screen, they ran pretty well.

ppppenguin: As Steve mentioned I am proposing this to help out other collectors. The price of approximately £322 is split about half for parts and half for hand assembly and test. I've built about a dozen so far by hand, and it's not a lot of fun (About 10 hours a piece under a microscope. ) I had the company that assembles our broadcast products quote it, and they also quoted hand assembly as tooling would be prohibative. As you know with these Xilinx parts, if we could get the volume up, and get it tooled, we could probably be selling them for less than £150, but I doubt there are that many crazies like us in the world to justify it
__________________
Aurora video standards converters: http://www.tech-retro.com/Aurora_Design/Video_Home.html
tubesrule is offline  
Old 10th Jul 2004, 9:56 am   #4
oldeurope
Retired Dormant Member
 
oldeurope's Avatar
 
Join Date: Dec 2003
Location: Solingen, Germany
Posts: 727
Default Re: New Multi-Standards Converter

Good morning all
Darryl, I watched the list of standards.
What is about the German French 441/25i
and French 819/25i
These are importand standards here in " old Europe " .
Of coures the 441 sets work with 405 too and vise
versa. But 819 would be interesting, because
my analog converter can not make 819 lines.

Kind regards
Darius
oldeurope is offline  
Old 10th Jul 2004, 11:03 am   #5
fableglade
Guest
 
Posts: n/a
Default Re: New Multi-Standards Converter

Good points Darius.

One other thought, I notice that the converter can output the current NTSC 525 line version it may also be well worth including the current British/PAL 625 line version as 625 line sets have beer around in England since the early 60's and even earlier in Europe and other countries I believe, so it people from the USA wanted to collect early 625 line sets it would enable them to generate a signal for these sets too in the same way that we over here could generate a signal for the American 525 line sets.

The addition of 819 lines is a very good idea, I have an 819 set which has been adjusted for 625, however I'd much prefer to have it running at 819 lines.
 
Old 10th Jul 2004, 6:53 pm   #6
tubesrule
Hexode
 
tubesrule's Avatar
 
Join Date: Jul 2004
Location: Michigan USA
Posts: 325
Default Re: New Multi-Standards Converter

Fableglade: The power grid was quite a confusing mess in the US and Canada into the 30's. There was no standardization even in a single community, and power could be 25Hz, 50Hz, 60Hz, or DC. By the 30's most big cities were standardized to 60Hz, but rural areas were still mixed into the 40's.

Fableglade/Darius: These are good suggestions for standards. The 441/25 should definitely be included, and I would think dropping the early prototype US standards like 120/24 and 240/24 for the early European 180/25 would be useful. (don't know how many of these sets survive, but for completeness..)
I have had suggestions for the 819 standard, but will need to look into this one to say for sure how it would be done. Right now all the output standards are smaller then the input video, so there are only downscalers in the converter right now. I think it would be a relatively simple matter to handle the 819, but it needs to be looked at.
__________________
Aurora video standards converters: http://www.tech-retro.com/Aurora_Design/Video_Home.html
tubesrule is offline  
Old 11th Jul 2004, 6:47 pm   #7
ppppenguin
Retired Dormant Member
 
ppppenguin's Avatar
 
Join Date: Dec 2003
Location: North London, UK.
Posts: 6,168
Default Re: New Multi-Standards Converter

Fortunately the SAA7113 decoder that Darryl uses at the input of his converter will accept almost any modern standard and deal with it well. It will lock to poor quality VCR output without problems. Since Darryl is using a framestore there should be no timebase errors at the output.

My view, which may be different from Darryl's, is that the output clock should be free running, unrelated to the input clock. This is true on the Domino converter and also on my own experimental design. (My design is never going into production, it's purely for my own amusement, using a spare card left over from a design for one of my clients. It could probably do 819 fairly easily though I have no real intention to do that.) The separate output clock ensures that the output timebase is completely stable regardless of input.

A further thought on decoder chips. I have used many of the Philips devices over the years. The SAA7113 is a decent design but does not have a luminance comb filter. I use the SAA7118 (which does have a comb) but I can't say I actually like it very much. And it's a ball grid package. For ultimate quality on composite inputs the comb may be worthwhile. I am looking at the latest parts from Analogue devices (ADV7181, 7183A and 7189). The last of these is expensive and not needed in most applications. I will most likely use the ADV7183A (note that this is very different from the earlier ADV7183) and hope to borrow an evaluation kit from the local AD distributor.

ppppenguin is offline  
Old 11th Jul 2004, 9:28 pm   #8
ppppenguin
Retired Dormant Member
 
ppppenguin's Avatar
 
Join Date: Dec 2003
Location: North London, UK.
Posts: 6,168
Default Re: New Multi-Standards Converter

Steve:

Although I do not have full design details of Darryl's converter I do know quite a lot about the Pineapple. Most converter designs have compromised filtering for simplicity. I suspect that Darryl has done a good job here. The input side will certainly have adequate bandwidth because it relies on the SAA7113 chip which is Ok in this department. I don't know about the output filtering.

The Pineapple has fairly simple filtering. The Dinosaur probably has slightly better filters. The Domino is limited partly by simple filtering but also by its 10MHz input sampling frequency. Darius's design (in the versions I have looked at closely) uses fairly low sampling frequencies for its line memories.
ppppenguin is offline  
Old 12th Jul 2004, 1:30 pm   #9
David_Robinson
Retired Dormant Member
 
David_Robinson's Avatar
 
Join Date: Feb 2004
Location: Nottingham, UK
Posts: 94
Default Re: New Multi-Standards Converter

This question of whether the output clock should be free-running, i.e. whether the standards converter should act as a frame synchroniser is interesting. There is another issue with a synchroniser that has not been mentioned. If there is an occasional skip or repeat frame this is not too serious, however if the input is a bit wobbly, such as an analogue VCR, it is possible to get many skips and repeats in a short period. This happens when the synchroniser is close to its maximum or minimum delay i.e. on the edge. Professional frame synchronisers overcome this by having hysteresis in the skip/repeat function, which means having slightly more than a frame of memory.

So a frame synchroniser is easy to do crudely but quite complicated to get right. Also of course there is the issue of audio/video delay errors when the video can be delayed by up to a frame. And while it is easy to get a dual-port RAM, or to get a big RAM, it is harder to source a big, dual-port RAM - except dedicated frame store chips which may well not support obsolete standards like 405. With line stores it's easier to get a PLD with enough internal RAM to do the job.

In my view it is only worth having a frame synchroniser when the output must be locked to an external reference.

Even my Sony 100Hz TV has a frame synchroniser. I really can't see the point. The video delay is not just large, but variable, and there is no audio delay compensation. And if I put a source such as a Digibox on the top shelf under the set, I get faint vertical bars scrolling across the screen. This is of course due to pickup of line flyback pulses from the TV line output stage, by the Digibox. :
David_Robinson is offline  
Old 12th Jul 2004, 2:47 pm   #10
tubesrule
Hexode
 
tubesrule's Avatar
 
Join Date: Jul 2004
Location: Michigan USA
Posts: 325
Default Re: New Multi-Standards Converter

Hi Darius: 1) Actually the converter already does this. The output is always generating a stable sync. When no video inout is connected, the output will free run. Also in this mode, there is a 32Mb FLASH memory on board for storing images. I usually put in a couple of different test cards, but the user can store anything they like. This way the output not only free runs with no input, but displays a stored image or an internal test pattern.

2) It would be possible to recontruct a 4:3 image from the 16:9 letterbox format. Of course the actual image resolution would be much less since so much information got discarded inthe letterbox process.

3) When outputing 405/25 or any other standard, you need to provide a standard PAL/SECAM input to the converter. All output standards are derrived form this input.

Hi David: The box in it's current implementation does provide hysteresis on the clock regeneration. The input clock could come from something as poor as a consumer VCR with a stretched tape, and the output clock would still be in spec. This is accomplished through algorithms in the fpga and an external PLL. When no video input is provided, the output clock free runs so a stable output is always provided.
Also in it's current form the converter has three full frames worth of dual port memory for a total of 1.5MB. As you mentioned true dual port memory of large sizes are not available, and even smaller ones are outragiously expensive. The dual port memory on the converter was constructed from standard computer cache RAM's with the fpga acting as the dual port control. Three full frames were needed for the worst case rate conversions, so the audio I/O handles resyncing the audio to match the video.
__________________
Aurora video standards converters: http://www.tech-retro.com/Aurora_Design/Video_Home.html
tubesrule is offline  
Old 29th Jul 2004, 4:19 pm   #11
ppppenguin
Retired Dormant Member
 
ppppenguin's Avatar
 
Join Date: Dec 2003
Location: North London, UK.
Posts: 6,168
Default Re: New Multi-Standards Converter

You execute the standard coder equations.

1. LPF the Cb and Cr components. Simple 1/4 + 1/2 +1/4 FIR is about right
2. Scale them according to the PAL or NTSC weighting factors. Now you have U and V
3. Multiply them by subcarrier (with 90 degree phase difference. And PAL switch if needed)
4. Add them together
5. Scale the Y.
6. Add Y to modulated C

Since the clock rate will be several times the Cb/Cr sample rate you can save a lot of multipliers.

Subcarrier is generated by an adder/accumulator and a sine lookup table. Reset the accumulator every 4 or 8 fields if you want absolutely correct SCH phase. 90 degree and 180 degree phase switching are simple to do ahead of the sine table.

Burst and sync edge shapes are pre-calculated and stored in small ROMs.

Burst is modulated along with Cb and Cr in the same multiplier(s). May be easier to do PAL burst modulation in just one multiplier rather than separate U and V components. This means phase shifting the subcarrier by 45 degrees. This is how I did my colour black.

Don't ask me how to do SECAM coding! I don't want to go there. Actually I do know a little about SECAM, having worked for Michael Cox Electronics many years ago. Mike was a great SECAM enthusiast.
ppppenguin is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 1:00 pm.


All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at https://www.vintage-radio.net/rules.html.
Failure to heed this warning may result in death or serious injury to yourself and/or others.


Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.